SQL/MX Release 2.0 Best Practices

Project Activities Guideline for Database Activities 9
Project Activities Guideline for Database Activities
This section lists activities that have been successfully employed at HP accounts. The list is intended as a
guide to acquiring the information and data necessary to implement a database solution successfully.
Assumptions include the use of a phased approach to limit the scope of the project and to implement
additional functionality in subsequent phases. The completion time for each task will vary based on the
customer’s experience with successfully implementing database solutions. The actual time will be a
function of the overall project’s complexity and scope.
Requirements Definition
• User interviews. Conduct interviews with representatives of user groups and business groups. The
customer’s project-team member who represents the interests of the group that will regularly use the
system should be present for all interviews. Allow 60 to 90 minutes for each interview. Schedule no
more than four interviews each day, leaving 30 minutes between interviews in which the interviewers
can compare notes.
• Document requirements. Encapsulate, combine, and consolidate the interview notes to produce a
draft of the requirements definition.
• User review. Review the draft requirements with the user groups. Take no more than 30 minutes with
each user group.
• Requirements completion. Produce completed requirements-definition documentation.
High-level Architecture and Design
Architecture
Obtain these information:
• Type of database
• Number of levels
• Source identification
• Data flow
• Change data capture
• Transformation and cleansing
• Interaction among components
• Network requirements
• Service-level agreements
• Volume of data
Logical Design
• Data analysis. Identify all data sources. Determine source data types, formats, ranges, and
validation. Obtain row and record counts, unique entry counts, data distribution, and second high
and low values.
• Logical database design. Define all entities and attributes. Identify all candidate keys.
• Data cleansing. Determine the correct format, ranges, and validation for data from disparate
sources. Determine the system of record for the data.
• Logical schema. Determine the data model that will represent the logical database.
• Data transformation. Determine the desired format, ranges, and validation for data to be loaded.